Create a Composite BitmapImage in WPF

Question

I have three BitmapImages that I would like to stitch together to create a composite image. The three images to be stitched together are aligned in the following way:

The images are of a type System.Windows.Media.Imaging.BitmapImage. I have looked at the following solution, but it uses System.Drawing.Graphics to perform stitching. I find it unintuitive to convert my BitmapImage to System.Drawing.Bitmap everytime I want to stich them together.

Is there a simple way to stitch three images of type System.Windows.Media.Imaging.BitmapImage together?

Use [CopyPixels](https://msdn.microsoft.com/en-us/library/ms616042.aspx) to get the pixel buffer of the source images, and [WritePixels](https://msdn.microsoft.com/en-us/library/cc490065.aspx) to put them into a WriteableBitmap. — Clemens, Jun 22 '15 at 23:36
@Clemens Yes, that was what I was thinking too. I was wondering if there was a much simpler way to do such a basic operation. I will do it and paste the code here. — siavashk, Jun 22 '15 at 23:41

Clemens · Accepted Answer · 2015-06-23T09:36:10.767

In addition to the options described in the other answer, the code below stitches three BitmapSource together into a single WriteableBitmap:

public BitmapSource StitchBitmaps(BitmapSource b1, BitmapSource b2, BitmapSource b3)
{
    if (b1.Format != b2.Format || b1.Format != b3.Format)
    {
        throw new ArgumentException("All input bitmaps must have the same pixel format");
    }

    var width = Math.Max(b1.PixelWidth, b2.PixelWidth + b3.PixelWidth);
    var height = b1.PixelHeight + Math.Max(b2.PixelHeight, b3.PixelHeight);
    var wb = new WriteableBitmap(width, height, 96, 96, b1.Format, null);
    var stride1 = (b1.PixelWidth * b1.Format.BitsPerPixel + 7) / 8;
    var stride2 = (b2.PixelWidth * b2.Format.BitsPerPixel + 7) / 8;
    var stride3 = (b3.PixelWidth * b3.Format.BitsPerPixel + 7) / 8;
    var size = b1.PixelHeight * stride1;
    size = Math.Max(size, b2.PixelHeight * stride2);
    size = Math.Max(size, b3.PixelHeight * stride3);

    var buffer = new byte[size];
    b1.CopyPixels(buffer, stride1, 0);
    wb.WritePixels(
        new Int32Rect(0, 0, b1.PixelWidth, b1.PixelHeight),
        buffer, stride1, 0);

    b2.CopyPixels(buffer, stride2, 0);
    wb.WritePixels(
        new Int32Rect(0, b1.PixelHeight, b2.PixelWidth, b2.PixelHeight),
        buffer, stride2, 0);

    b3.CopyPixels(buffer, stride3, 0);
    wb.WritePixels(
        new Int32Rect(b2.PixelWidth, b1.PixelHeight, b3.PixelWidth, b3.PixelHeight),
        buffer, stride3, 0);

    return wb;
}

Thanks! This works like a charm. I have one question though, what does the magic number 7 do when you are calculating the stride? — siavashk, Jun 23 '15 at 20:13
It is necessary to calculate the correct number of bytes-per-scan-line (i.e. the stride) when bits-per-pixel is not an integer multiple of 8. The common bitmap formats have either 24 or 32 bits-per-pixel, so you could also calculate stride as width * bpp / 8. — Clemens, Jun 23 '15 at 21:28

score 2 · Answer 2 · answered Jun 23 '15 at 04:28

It really depends on what output type you want. Here are some options.

DrawingVisual

If you just need to render them to a visual, you could use a DrawingVisual and render the three images. You could then render the visual in various ways depending on your use case (for example, using a VisualBrush).

Rect bounds = new Rect(0.0, 0.0, 400.0, 300.0);
DrawingVisual visual = new DrawingVisual();
DrawingContext dc = visual.RenderOpen();
dc.DrawImage(mImage, new Rect(bounds.X, bounds.Y, bounds.Width, bounds.Height * 0.5));
dc.DrawImage(mImage, new Rect(bounds.X, bounds.Y + bounds.Height * 0.5, bounds.Width * 0.75, bounds.Height * 0.5));
dc.DrawImage(mImage, new Rect(bounds.X + bounds.Width * 0.75, bounds.Y + bounds.Height * 0.5, bounds.Width * 0.25, bounds.Height * 0.5));
dc.Close();

Custom Element

If you need an element that you can place in your UI directly, you can make a custom element that extends FrameworkElement.

class CustomElement : FrameworkElement
{
    public ImageSource Image1
    {
        get { return (ImageSource)GetValue(Image1Property); }
        set { SetValue(Image1Property, value); }
    }
    public static readonly DependencyProperty Image1Property = DependencyProperty.Register("Image1", typeof(ImageSource), typeof(CustomElement),
        new FrameworkPropertyMetadata((ImageSource)null, FrameworkPropertyMetadataOptions.AffectsRender));

    public ImageSource Image2
    {
        get { return (ImageSource)GetValue(Image2Property); }
        set { SetValue(Image2Property, value); }
    }
    public static readonly DependencyProperty Image2Property = DependencyProperty.Register("Image2", typeof(ImageSource), typeof(CustomElement),
        new FrameworkPropertyMetadata((ImageSource)null, FrameworkPropertyMetadataOptions.AffectsRender));

    public ImageSource Image3
    {
        get { return (ImageSource)GetValue(Image3Property); }
        set { SetValue(Image3Property, value); }
    }
    public static readonly DependencyProperty Image3Property = DependencyProperty.Register("Image3", typeof(ImageSource), typeof(CustomElement),
        new FrameworkPropertyMetadata((ImageSource)null, FrameworkPropertyMetadataOptions.AffectsRender));

    protected override void OnRender(DrawingContext drawingContext)
    {
        drawingContext.DrawImage(Image1, new Rect(0.0, 0.0, ActualWidth, ActualHeight * 0.5));
        drawingContext.DrawImage(Image2, new Rect(0.0, ActualHeight * 0.5, ActualWidth * 0.75, ActualHeight * 0.5));
        drawingContext.DrawImage(Image3, new Rect(ActualWidth * 0.75, ActualHeight * 0.5, ActualWidth * 0.25, ActualHeight * 0.5));
    }
}

You could then use it like this:

<local:CustomElement
    Image1="{Binding SomeImage}"
    Image2="{Binding SomeOtherImage}"
    Image3="http://stackoverflow.com/favicon.ico" />

Images in a Grid

You always have the option of putting three Image controls in a Grid.

<Grid>
    <Grid.RowDefinitions>
        <RowDefinition Height="*" />
        <RowDefinition Height="*" />
    </Grid.RowDefinitions>
    <Grid.ColumnDefinitions>
        <ColumnDefinition Width="3*" />
        <ColumnDefinition Width="*" />
    </Grid.ColumnDefinitions>
    <Image
        Grid.ColumnSpan="2"
        Source="{Binding SomeImage}" />
    <Image
        Grid.Row="1"
        Source="{Binding SomeOtherImage}" />
    <Image
        Grid.Row="1"
        Grid.Column="1"
        Source="http://stackexchange.com/favicon.ico" />
</Grid>

Creating an ImageSource

If you need the three images combined into a single ImageSource for some reason, you could render into a DrawingVisual (mentioned above) and then render that visual into a RenderTargetBitmap.

Create a Composite BitmapImage in WPF

2 Answers2

Linked

Related